Automated Discrimination of Lightning Stepped Leader Stage from the Power Spectrum Density of the Related Electric Field Recordings

This paper is aimed to show how to detect and identify the lightning stepped leader stage from the power spectrum density spectrograms (PSD) of related lightning electric field records. Such electric field records were obtained from the local lightning detection system operating in the Rzeszow region. This system consists of two slow and fast electric field sensors working in the ELF and ELF-MF frequency range, respectively. Additionally, the development of the corresponding lightning flash channel can be recorded by the digital high speed video camera. The frequency bandwidth of the electric field fast antenna enable us to recognize properly all main components of lightning discharge observed from the single measuring station. On the other hand, the lightning stepped leader stage of the considered cloud-to-ground flash can be determined from the modified power density spectrograms of related lightning electric field records. Such PSD spectrograms have been integrated in respect to the frequency domain. The only-time-dependent PSD spectrograms have been recalculated by the special computer procedure called the fill factor estimation. This factor has been computed in respect to the following return stroke electric field signature. Next, the leader beginning timestamp has been determined for the particular value of the fill factor. The obtained duration of the lightning leader stage enable us to estimate the lightning channel height for simplified assumptions. The comparison of the downward negative leader model taken from the literature with our lightning electric field records is also shown. The presented computer algorithm can be used for the different kind of lightning electric records. The high efficiency of such lightning leader discrimination has been obtained especially in the cases of close lightning flash incidents when electrostatic field component of the leader stage dominates the total electric field change of the considered lightning stroke. Therefore, this method may be implemented in future by the different kind of lightning detection and location systems without a significant loss of computing power and with low costs.